Synthetic Minor Determination and/or Netting Determination 04
MS312 rapid method for determination of 228RA in water samples
Transcript of MS312 rapid method for determination of 228RA in water samples
1 23
Journal of Radioanalytical andNuclear ChemistryAn International Journal Dealing withAll Aspects and Applications of NuclearChemistry ISSN 0236-5731 J Radioanal Nucl ChemDOI 10.1007/s10967-012-2257-1
Rapid method for determination of 228Rain water samples
Sherrod L. Maxwell, Brian K. Culligan,Robin C. Utsey, Daniel R. McAlister &E. Philip Horwitz
1 23
Your article is protected by copyright and
all rights are held exclusively by Akadémiai
Kiadó, Budapest, Hungary. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.
Rapid method for determination of 228Ra in water samples
Sherrod L. Maxwell • Brian K. Culligan •
Robin C. Utsey • Daniel R. McAlister •
E. Philip Horwitz
Received: 5 September 2012
� Akademiai Kiado, Budapest, Hungary 2012
Abstract A new rapid method for the determination of228Ra in natural water samples has been developed at the
SRNL/EBL (Savannah River National Lab/Environmental
Bioassay Laboratory) that can be used for emergency
response or routine samples. While gamma spectrometry
can be employed with sufficient detection limits to deter-
mine 228Ra in solid samples (via 228Ac), radiochemical
methods that employ gas flow proportional counting tech-
niques typically provide lower minimal detectable activity
levels for the determination of 228Ra in water samples.
Most radiochemical methods for 228Ra collect and purify228Ra and allow for 228Ac daughter ingrowth for *36 h.
In this new SRNL/EBL approach, 228Ac is collected and
purified from the water sample without waiting to eliminate
this delay. The sample preparation requires only about 4 h
so that 228Ra assay results on water samples can be
achieved in \6 h. The method uses a rapid calcium car-
bonate precipitation enhanced with a small amount of
phosphate added to enhance chemical yields (typically
[90 %), followed by rapid cation exchange removal of
calcium. Lead, bismuth, uranium, thorium and protactin-
ium isotopes are also removed by the cation exchange
separation. 228Ac is eluted from the cation resin directly
onto a DGA Resin cartridge attached to the bottom of the
cation column to purify 228Ac. DGA Resin also removes
lead and bismuth isotopes, along with Sr isotopes and 90Y.
La is used to determine 228Ac chemical yield via ICP-MS,
but 133Ba can also be used instead if ICP-MS assay is not
available. Unlike some older methods, no lead or strontium
holdback carriers or continual readjustment of sample pH
is required.
Keywords Rapid � 228Ra � 228Ac � DGA Resin �Natural waters
Introduction
The measurement of radium isotopes in natural waters is
important in the study of hydrospheric and geochemical
processes, as well as for public health reasons. Large quan-
tities of radium isotopes can be produced or accumulated as
waste or by products from various industries, including
phosphate mining and the oil and gas industry. Radium tends
to accumulate in the bones of mammals due to chemical
behavior similar to other alkaline earth metals (calcium,
barium and strontium). As a result, the consumption of food
and water containing radium isotopes by the general public,
along with their short-lived radioactive progeny with high
specific activities, can increase significantly the internal
radiation dose of individuals [1]. Due to this potential tox-
icity of radium isotopes and their importance in geochemical
studies related to mixing processes, the measurement of226Ra and 228Ra in natural waters continues to be very
important [2, 3]. Recently, new rapid methods for 226Ra
developed in the Savannah River National Lab/Environ-
mental Bioassay Lab (SRNL/EBL, Aiken, SC, USA) were
reported that provide rapid, effective assay of 226Ra in a
variety of environmental matrices [4]. 225Ra (217At), was
used as a yield tracer in this rapid alpha spectrometry
method, eliminating potential problems associated with133Ba tracer and Ra isotopes behaving differently.
S. L. Maxwell (&) � B. K. Culligan � R. C. Utsey
Savannah River National Laboratory, Building 735-B, Aiken,
SC 29808, USA
e-mail: [email protected]
D. R. McAlister � E. P. Horwitz
PG Research Foundation, Inc., Lisle, IL 60532, USA
123
J Radioanal Nucl Chem
DOI 10.1007/s10967-012-2257-1
Author's personal copy
Many 228Ra methods utilize liquid–liquid extraction
using bis (2-ethylhexyl) phosphoric acid (HDEHP) to
collect and purify 228Ac to measure 228Ra indirectly (after
36 h ingrowth). These solvent extraction techniques can be
tedious and generate large amounts of liquid solvent waste.
Burnett used Ln-Resin� (Eichrom Technologies, Lisle, IL,
USA), which is a resin coated with HDEHP, to separate228Ac after a barium sulfate precipitation was used to
collect radium from natural water samples. In this method,
the barium sulfate precipitate had to be converted to bar-
ium carbonate in a heat bath for several hours. [5] Man-
ganese dioxide (MnO2)—coated fibers have been used for
years for oceanographic analyses to collect radium for
analysis. [3].
Nour et al. [6] also used MnO2 precipitation and
Diphonix Resin (Eichrom Technologies) to collect 228Ra
and 228Ac. Stable strontium was added as a holdback car-
rier to try to minimize interference from Sr isotopes. 90Sr,
if not removed, will quickly regenerate 90Y, which will
follow 228Ac on Diphonix Resin in this method. After
ingrowth, 228Ac was eluted with 1 M hydroxyethane-1,1
diphosphonic acid (HEDPA) and determined by liquid
scintillation counting. The HEDPA was used because
Ac(III) is very difficult to elute from Diphonix Resin. The
use of HEDPA eluent typically precludes the use of gas
flow proportional counting and subsequent separations, but
it can be used by liquid scintillation or Cherenkov count-
ing. Both techniques, however, typically have higher
minimum detectable activity (MDA)’s than gas flow pro-
portional counting. Aleissa et al. [7] also used the separa-
tion chemistry reported by Nour et al. but applied
Cherenkov counting to determine 228Ac instead of liquid
scintillation. Sodium salicylate was added to increase228Ac efficiencies by Cherenkov counting from 10 to 38 %.
A combined 226Ra/228Ra method was reported by the
SRNL/EBL lab that utilized MnO2 Resin (Eichrom Tech-
nologies) to collect radium from water samples. The 228Ra
method required a 30 h ingrowth period for the ingrowth of228Ac, prior to separation using DGA Resin. The method used133Ba as a yield monitor for both the 226Ra and 228Ra mea-
surements [8]. While there were some advantages to using
MnO2 Resin, the previous work did not have a 228Ac yield
monitor that was also retained and eluted from the DGA
Resin. The method also required an ingrowth period for 228Ac.
Based on a survey of the scientific literature, there still
seemed to be a need for an improved rapid 228Ra method,
especially if the there is a desire to collect the 228Ac from
water samples without waiting for ingrowth. Since MnO2
Resin or MnO2 precipitation may not collect 228Ac quan-
titatively under certain sample matrix conditions, another
separation approach was investigated.
A new rapid method for the determination of 228Ra in
natural water samples has been developed at the SRNL/
EBL Laboratory that can be used for emergency response
or routine samples. While gamma spectrometry can usually
be employed with sufficient detection limits to determine228Ra in solid samples (via 228Ac), radiochemical methods
that employ gas flow proportional counting techniques
typically provide lower MDA levels for 228Ra assay for a
water sample matrix. Most radiochemical methods for228Ra collect and purify 228Ra and allow for 228Ac daughter
ingrowth for *36 h. In this new SRNL/EBL approach,228Ac is collected and purified from the water sample
without waiting to eliminate this delay. The sample prep-
aration requires only about 4 h so that 228Ra assay results
on water samples can be achieved in \6 h. This new
approach for the assay of 228Ra offers an improved yield
measurement for 228Ac and a rapid option to collect 228Ac
without waiting for ingrowth.
The method uses a rapid calcium carbonate precipitation
enhanced with a small amount of phosphate to enhance
chemical yields (typically[90 %), followed by rapid cation
exchange removal of calcium. Lead, bismuth, uranium, tho-
rium and protactinium isotopes are also removed by the cat-
ion exchange separation [9]. 228Ac is eluted from cation resin
directly onto a DGA Resin cartridge attached to the cation
column. DGA Resin is used to purify 228Ac, also removing
lead and bismuth isotopes, along with Sr isotopes and 90Y. La
is used to determine 228Ac chemical yield via inductively-
coupled plasma-mass spectrometry (ICP-MS), but 133Ba can
also be used instead if ICP-MS assay is not available.
Unlike other methods, the method outlined in this work
does not require any liquid solvents such as HDEHP, does
not need a carbonate conversion of barium sulfate for hours
in a heat bath, does not require precise load solution acidity
adjustments with Ln Resin, and does not require waiting
36 h for 228Ac ingrowth. However, 228Ac ingrowth is an
option if desired. Unlike some older methods, no holdback
carriers or continual readjustment of sample pH is required.
If 139,141Ce, 147Nd and 140La radionuclides are present
due to a mixed fission product release, one can collect the
purified 228Ra from the sample after rapid removal of these
lanthanides using DGA Resin and reprocess the samples
through DGA Resin again after ingrowth of 228Ac.
The method ruggedness, as indicated by high chemical
yields and effective removal of interferences, including
high levels of 90Sr, makes this method useful for envi-
ronmental laboratories.
Experimental
Reagents
The resins employed in this work are Cation Resin
(50 W-X8, Hydrogen form, 200–400 mesh) and DGA
S. L. Maxwell et al.
123
Author's personal copy
Resin (N,N,N0,N0 tetraoctyldiglycolamide, available from
Eichrom Technologies, LLC (Lisle, IL, USA). Nitric and
hydrofluoric acids were prepared from reagent-grade acids
(Fisher Scientific, Inc.). All water was obtained from a Milli-
Q2TM water purification system. All other materials were
ACS reagent grade. The radiochemical isotope 228Ra was
obtained from Eckert and Ziegler/Analytics, Inc. (Atlanta,
GA, USA) and diluted to approximately 0.37 Bq m/L. 90Sr,226Ra, and 238 U standards were also obtained from Eckert
and Ziegler/Analytics, Inc.
Procedures
Column preparation
Cation exchange resin (Eichrom 50WX8, 200–400 mesh)
was obtained as bulk resin and columns were prepared by
weighing out the resin amounts and adding to in large ion
exchange column reservoirs (Environmental Express,
Mount Pleasant, SC, USA). DGA Resin cartridges contain-
ing 2 mL of each resin were obtained from Eichrom Tech-
nologies, (Lisle, IL). Small particle size (50–100 micron)
resin was employed, along with a vacuum extraction system
(Eichrom Technologies) that will handle 24 samples at a
time. Flow rates of *1–2 mL/min were typically used.
The separation utilizes removal of interferences using
both cation exchange resin and DGA Resin. The cation
resin effectively removes large amounts of calcium used to
precipitate the 228Ra, 228Ac and La ions from water sam-
ples, as well as other interferences. While calcium, stron-
tium and lead ions are all strongly retained on DGA Resin
in 3 M HNO3, they are all effectively removed using much
stronger nitric acid. Since there is much less retention of
Ca, Sr and Pb ions in 7 M HNO3, this acid concentration
was selected to elute La, Ac and Ra from the cation resin
through DGA Resin [10]. Figure 1 shows how Ac3?
retention on DGA Resin is still significant (k0[ 100) in
7 M HNO3. This nitric acid concentration was selected to
remove Ca, Sr and Pb ions, while retaining Ac and La ions
on DGA Resin. The k0 values were measured using the
batch extraction procedure and calculations described in
reference [11].
Figure 1 also shows how Ac3? can be eluted from DGA
Resin using 2 M HCl. Under these conditions, Y, Po and Bi
ions (as well as heavier lanthanide isotopes such as Eu, Gd)
will be retained. Figure 2 shows the k0 retention values are
very low for La in 2 M HCl, while Y is retained.
Sample preparation
Replicate sample aliquots of various water samples were
added to 1 liter glass beakers. Tap water and well water
samples were spiked with varying levels of 228Ra.
Figure 3 shows the rapid precipitation steps that can be
used for aqueous environmental samples to preconcentrate
the 228Ra from a water matrix. Water sample aliquots of 1
liter were acidified to *pH 2 in glass beakers using nitric
acid. To each test batch of replicate water samples, varying
levels of 228Ra were added to each replicate sample. Water
sample batches were spiked with 228Ra at 177.2 mBq/L
(4.78 pCi/L), 354.5 mBq/L (9.57 pCi/L), and 1,046 mBq/L
(28.2 pCi/L) respectively. A blank replicate was also ana-
lyzed with each batch so that the 228Ra results could be
corrected for the native content of 228Ra. Lanthanum (La)
tracer (1 mg) was added to each sample. Barium carrier
(5 mg) was added to each sample to enhance precipitation
efficiency. Three milliliters of 1.25 M calcium nitrate
(150 mg Ca) and 1 mL of 3.2 M ammonium hydrogen
phosphate were added to each water sample. Thirty milli-
liters of concentrated ammonium hydroxide (14.5 M) were
added to each sample beaker and each solution was mixed.
Thirty milliliters of 2 M sodium carbonate were added to
each sample beaker. Each sample was stirred and placed on
a hot plate on high heat for about 30 min or until very hot
to reduce the impact of any dissolved CO2 in the samples
and to reduce solubility.
The beakers were removed from the hot plate and the
precipitate was allowed to settle. The supernatant was
poured off down to about 200 mL volume so that the
remaining sample and solids could be transferred to
225 mL centrifuge tubes. The tubes were centrifuged at
3,500 rpm for 5 min and the supernatant was discarded.
The remaining solids were dissolved in 10 mL of 1.5 M
HCl, and transferred to a 50 mL centrifuge tube. The
225 mL tube was rinsed well with 7 mL of 1.0 M HCl and
this rinse solution was added to each dissolved sample.
If any residual solids remained, they were rinsed well by
mixing with 5 mL 1.5 M HCl, and centrifuging to remove
any residual solids. This rinse was added to the original
sample solution.
Samples were loaded to columns containing 5.0 g of
Cation Resin (200–400 mesh). The bed height of the col-
umns was about 4 cm and each column was conditioned
with *25 mL of deionized water and 10 mL of 0.5 M
HCl. Gravity flow was typically sufficient to achieve a flow
rate of *1 drop/s, however vacuum was applied if needed.
Column separation
Figure 4 shows the rapid column separation sequence used.
The Ra, Ac and La were retained on cation resin (5 g) and
calcium was removed by rinsing with 2.75 M HCl-0.02 M
HF at *1 drop/s. HF was used in this HCl rinse step to
help stabilize any Pa5? ions that might be present, so they
would pass through the cation resin without adhering to the
plastic surfaces such as the polyethylene frits in the cation
Rapid method for determination
123
Author's personal copy
column. A 2 mL DGA Resin cartridge was placed below
each cation resin column and 228Ac and La were eluted
from the cation resin with 25 mL 7 M HNO3 onto DGA
Resin at *1 drop/s, while Ra passed through. The sepa-
ration time was taken as the mid-point of this elution step
to accurately determine and correct for any decay of 228Ac
in the following separation and counting steps. The cation
resin column was removed and DGA Resin alone was
rinsed with 5 mL 7 M HNO3 to further enhance the
removal of Sr and Pb isotopes.228Ac and La were eluted with 19 mL 2 M HCl at *1
drop/s. The eluent volumes were adjusted to exactly 20 mL
volume with 2 M HCl and mixed well. One hundred
microliters of each sample were transferred to 50 mL
10-2 10-1 100 10110-1
100
101
102
103
104
10-2 10-1 100 10110-1
100
101
102
103
104
k' f
orA
c(II
I)
[HNO3] [HCl]
50-100 µm resin, 1 h contact time, 22(2) °C
k' 225Ac vs. [HNO3] or HCl on DGA Resin, Normal Fig. 1 Retention of Ac(III) on
DGA resin courtesy PG
Research Foundation, Lisle, IL,
USA
1
100
101
102
103
0.70.5 432
Dy
Nd
Tb
Ho
GdEuSm
50-100 µm, 2 h, 21(1) ºCk' on DGA Resin vs HCl
k'
[HCl], M
Pm
1
100
101
102
103
0.70.5 432
Cm
Nd Pr
Am
CeLa
Sc
Y
Sm
50-100 µm, 2 h, 21(1)ºCk' on DGA Resin vs HCl
k'
[HCl], M
Pm
Fig. 2 Retention of lanthanides
and yttrium on DGA resin in
HCl courtesy PG Research
Foundation, Lisle, IL, USA
S. L. Maxwell et al.
123
Author's personal copy
centrifuge tubes for dilution to 50 mL and ICP-MS mea-
surement of La (*100 ng/mL).
Microprecipitation
To each final purified solution containing 20 mL 2 M HCl,
100 lg cerium and 3 mL 28 M HF were added. While the
rare earth La was already present in the purified solution,
Ce was included in the method in case La tracer was
inadvertently omitted or to allow for the use of 133Ba tracer
instead. After mixing well, the tubes were allowed to sit for
*15 min. The solutions were filtered onto 0.1 micron
25 mm polypropylene filters (Resolve �-Filter-Eichrom
Technologies), rinsing the filters with 95 % ethanol. The
filters were dried under a heat lamp and counted by gas
flow proportional counting.
Apparatus
Polycarbonate vacuum boxes with 24 positions and a rack
to hold 50 mL plastic tubes were used to perform the
column separations.
The 228Ac samples are counted simultaneously using a
thin (Mylar) windowed Tennelec LB 4100 Gas Flow Pro-
portional Counter system. A high detection efficiency
([50 %) and low detector background (\2 cpm) results in
detection levels typically\1 pCi/L for 60 min count times.
Results are decay corrected from the midpoint of the count
time back to the midpoint of the elution time.
A Perkin Elmer Elan DRC-e (using standard ICP-MS
mode) was used to perform the ICP-MS measurements.
Instrument operating conditions are shown in Table 1.
Results and discussion
Table 2 shows the individual results for the determination
of 228Ra (spiked at 177.2 mBq/L) in seven one liter tap
water samples using this rapid separation method and gas
flow proportional counting. The 228Ac/228Ra results were
corrected for La tracer yield using ICP-MS. The average228Ra result for the one liter tap water samples was
177.5 mBq/L, with a 0.2 % bias and 1 SD (standard
deviation) of 20.5 mBq/L. The measured values were
corrected for 21.5 mBq 228Ra found in the unspiked water
sample. The high La (228Ac) tracer recoveries and excellent
results for the analytes versus known values indicate the
ruggedness of the sample preparation and measurement
Fig. 3 Rapid 228Ra sample preparation for water samples
Fig. 4 Rapid 228Ra column separation for water samples
Rapid method for determination
123
Author's personal copy
steps for the water samples. The average tracer recovery
for La was 94.1 % ± 2.3 % at 1 SD.
Table 3 shows the individual results for the determination
of 228Ra (spiked at 354.5 mBq/L) in seven one liter tap water
samples using this rapid separation method and gas flow
proportional counting. The 228Ac/228Ra results were corrected
for La tracer yield using ICP-MS. The average 228Ra result for
the one liter tap water samples was 347.2 mBq/L, with a
-2.1 % bias and 1 SD (standard deviation) of 24.8 mBq/L.
The measured values were corrected for 21.5 mBq 228Ra
found in the unspiked water sample. The La (228Ac) tracer
recoveries were very good and very consistent. The average
tracer recovery for La was 91.8 % ± 1.2 % at 1SD.
Table 4 shows the individual results for the determina-
tion of 228Ra (spiked at 1,046 mBq/L) in six one liter
groundwater samples using this rapid separation method
and gas flow proportional counting. The 228Ac/228 Ra
results were corrected for La tracer yield using ICP-MS.
The average 228Ra result for the 1 L tap water samples was
1,008.3 mBq/L, with a -3.6 % bias and 1 SD (standard
deviation) of 27.8 mBq/L. The measured values were
corrected for 174.2 mBq 228Ra found in the unspiked water
sample. The average tracer recovery for La (228Ac) was
95.3 % ± 0.9 % at 1 SD.
Table 5 shows the individual results for the determination
of 228Ra in one liter water samples with varying levels of
interferences added. 90Sr/90Y at 5.92 Bq and 29.6 Bq levels
respectively were added to separate sets of 4 water samples
spiked with 1.046 Bq 228Ra. In addition, 3.7 Bq 238U and
4.81 Bq 226Ra were added to 4 sets of water samples also
spiked with 1.046 Bq 228Ra. The overall 228Ra recoveries are
*95 % for each set of samples spiked with interferences, with
no significant positive bias in the 228Ra results for these ratios
of 90Sr/90Y, 238U and 226Ra added. It was particularly
encouraging to see no apparent Sr/Y carryover into the 228Ac
fraction despite the addition of 800 pCi (29.6 Bq), considering
the difficulties that some 228Ra methods have with interfer-
ence from Sr/Y. The removal of 226Ra progeny such as 214Pb
and 214Bi daughters was also confirmed, despite adding226Ra/228Ra ratios of almost 5–1. A more precise decontam-
ination study can be performed using water blanks, but con-
sidering the typical Ra, U and Sr levels and ratios in natural
waters this test was still very useful, indicating excellent
removal of typical interferences.
Table 6 shows the results from spiked water blanks with
1,000 pCi/L (37,000 mBq/L) of 90Sr/90Y added, counted
with a 10 h count time. The average blank activity of 0.244
pCi/L (9.01 mBq/L) was very close to the average MDA of
0.202 pCi/L (7.46 mBq/L). While the blank results may
have a slight positive bias relative to zero, this bias in the
blank values is not considered to be significant from a
practical standpoint. A decontamination factor was calcu-
lated for each blank measurement by dividing the initial
1,000 pCi (37,000 mBq) of 90Sr/90Y added by the final
blank measurement. The average decontamination factor
for 90Sr/90Y was 4,230 (1SD = 797), indicating an average90Sr/90Y removal of 99.97 %.
The MDA for the actinide isotopes by alpha spectrometry
were calculated according to equations prescribed by Currie:
[11].
MDA ¼ ½3þ 4:65p
B�= CT � R �W � Eff � 0:060ð Þ
where B = total background counts, = BKG (rate) * BKG
count time, CT = sample count time (min), R = chemical
recovery, W = sample aliquot (L), Eff = detector effi-
ciency, 0.060 = conversion from dpm to mBq
In low-level counting, where a zero background count is
common, the constant 3 is used to prevent an excessively
high false positive rate.
The MDA for the 228Ra results can be adjusted as
needed, depending on the sample aliquot and count time.
This method provides a typical MDA of *18.5 mBq/L
(0.5 pCi/L) for a 90 min count time for a 1 L sample water
sample. Longer count times can be used to lower MDA
levels as needed.
Table 1 Operating conditions for Perkin Elmer Elan DRC-e
Plasma conditions
RF Power 1,400 W
Torch depth 5.5 mm
Plasma gas 15 L/min
Carrier gas 1 L/min
Nebulizer gas 0.98 L/min
Sample pump 5 rps
Ion lens/quadrupole
E1 lens voltage 6.25 V
E1 lens slope 0.0165
E1 lens intercept 4.413
Cell path voltage -12 CPV
Cell Rod offset -17 V
Q-pole rod offset -4 V
Detector
Discriminator 17 V
Analog HV -1,550 V
Pulse HV 900 V
Typical tune
Counts [300,000 cps In-115 at 10 lg/L
RSD % \5 %
Oxide 156/140 \5 %
Background \10 cps at mass 220; \ 10 cps at mass 8.5
(vacant mass -noise detection only)
Resolution 0.60–0.80 amu at 10 % peak height
Data acquisition
Integration 1,000 ms dwell time 50 ms
Replicates 3 with 20 sweeps/reading
S. L. Maxwell et al.
123
Author's personal copy
New resin cartridges were used for each analysis to
minimize any chance of cross-contamination of samples or
unexpected degradation of performance, which can occur
over time and may be different than the anticipated reuse
rate depending on real world sample matrix variation.
Some laboratories, however, have had success reusing
resins. It is anticipated that DGA Resin can potentially be
reused after rinsing the resin cartridges with 0.25 M HCl.
The use of stable La as a tracer for 228Ac provided a
very effective yield measurement. The rare earth fluoride
sample source mount was effective, and does not require
any consideration of hydroscopic impacts on a carrier
precipitate, such as yttrium oxalate, for yield As an alter-
native to La by ICP-MS, 133Ba can be added and used to
determine yield. The 7 M HNO3 load solution can be
collected after it passes through DGA Resin and counted
by gamma spectrometry.
If sediment is present in the water samples, the sediment
can be filtered on a glass fiber filter, for example, and a
rapid sodium hydroxide fusion can be applied [4]. The
fusion matrix can be dissolved in water and added back to
the 1 liter water sample if a combined water/sediment228Ra result is desired. If separate assays are preferred, the
same preconcentration chemistry can be applied. The cal-
cium carbonate/phosphate precipitation steps can be used
after transferring the alkaline fusion matrix to a 225 mL
centrifuge tube collect the 228Ra/228Ac in the sediment,
allowing for the hydroxide added from the fusion material.
The method assumes 228Ra/228Ac equilibrium, a rea-
sonable assumption especially 36 h after the water sample
is preserved by adjusting the acidity to pH 2. Water sam-
ples are often preserved at the field collection point,
however, if not, the samples may be held 36 h after
adjusting to pH 2 in the lab, or alternately the sample
precipitate may be held 36 h to allow ingrowth of 228Ac as
well. It is likely, however, that water samples, unless they
contain significant amount of sediment, are in equilibrium
prior to sample pH adjustment.
While it would seem the rapid 226Ra [4] and rapid 228Ra
methods recently developed by the SRNL/EBL lab could
Table 2 228Ra results for water samples (177 mBq/L level)
Sample
ID
La
(Ac)
Yield
(%)
228Ra
reference
value
(pCi/L)
228Ra
reference
value
(mBq/L)
228Ra
measured
value
(mBq/L)
Difference
(%)
1 93.4 4.79 177.2 160.5 -9.47
2 92.1 4.79 177.2 169.2 -4.51
3 95.3 4.79 177.2 167.4 -5.54
4 97.8 4.79 177.2 193.1 8.98
5 91.6 4.79 177.2 215.7 21.72
6 94.4 4.79 177.2 158.4 -10.61
7 95.6 4.79 177.2 178.3 0.60
Avg 94.3 177.5 0.2
SD 2.2 20.5
% RSD 2.3 11.6
90 min count time
1 L sample aliquot
La (Ac) yield by ICP-MS
Measured values corrected for 21.5 mBq per liter native 228Ra content
Table 3 228Ra results for water samples (354.5 mBq/L level)
Sample
ID
La
(Ac)
yield
(%)
228Ra
Reference
value
(pCi/L)
228Ra
Reference
value228
(mBq/L)
Ra
Measured
value
(mBq/L)
Difference
(%)
1 92.0 9.58 354.5 331.7 -6.4
2 93.1 9.58 354.5 380.9 7.5
3 93.1 9.58 354.5 309.2 -12.8
4 91.1 9.58 354.5 356.1 0.5
5 91.6 9.58 354.5 329.4 -7.1
6 90.1 9.58 354.5 367.0 3.6
7 94.0 9.58 354.5 355.9 0.4
Avg 92.1 347.2 -2.1
SD 1.3 24.8
% RSD 1.5 7.1
60 min count time
1 L sample aliquot
La (Ac) yield by ICP-MS
Measured values corrected for 21.5 mBq per liter native 228Ra content
Table 4 228Ra results for ground water samples (1,046 mBq/L level)
Sample
ID
La
(Ac)
Yield
(%)
228Ra
reference
value
(pCi/L)
228Ra
reference
value
(mBq/L)
228Ra
measured
value
(mBq/L)
Difference
(%)
1 94.9 28.28 1,046.4 1,033.6 -1.2
2 95.7 28.28 1,046.4 1,045.4 -0.1
3 96.2 28.28 1,046.4 1,017.1 -2.8
4 94.5 28.28 1,046.4 992.0 -5.2
5 96.4 28.28 1,046.4 982.5 -6.1
6 94.3 28.28 1,046.4 979.3 -6.4
Avg 95.3 1,008.3 -3.6
SD 0.9 27.8
% RSD 0.9 2.8
60 min count time
1 liter sample aliquot
La (Ac) yield by ICP-MS
Measured values corrected for 174.2 mBq per liter native 228Ra
content
Rapid method for determination
123
Author's personal copy
be combined into a single method, this would likely pre-
clude the use of 225Ra as a tracer, due to beta-emitting
progeny in the 225Ra decay chain. For labs that elected to
use 133Ba instead, combining these two approaches could
very well be a viable option, with associated cost benefits.
Conclusions
A new rapid method to determine 228Ra in natural water
samples has been developed that allows the separation of228Ra with high chemical yields and effective removal of
interferences. The new method uses a rapid calcium car-
bonate precipitation enhanced with a small amount of
phosphate to enhance chemical yields (typically [90 %),
followed by rapid cation exchange removal of calcium and
other interferences. Lead, bismuth, uranium, thorium and
protactinium isotopes are also removed by the cation
exchange separation. 228Ac is eluted from cation resin
directly onto a DGA Resin cartridge attached to the cation
resin column.to purify 228Ac. DGA Resin also removes
lead and bismuth isotopes, along with Sr isotopes and 90Y.
La is used to determine 228Ac chemical yield via induc-
tively-coupled plasma-mass spectrometry (ICP-MS), but
133Ba can also be used instead if ICP-MS assay is not
available.
Acknowledgments This work was performed under the auspices of
the Department of Energy, DOE Contract No. DE-AC09-96SR18500.
The authors wish to acknowledge Staci Britt, Jack Herrington and
Becky Chavous for their assistance with this work.
References
1. Jia G, Torri G (2007) Estimation of radiation doses to members of
the public in Italy from intakes of some important naturally
occurring radionuclides in drinking water (238U, 234U, 235U,226Ra, 228Ra, 224Ra and 210Po. Appl Radiat Isot 65:849–857
2. Jia G, Jia J (2012) Determination of radium isotopes in envi-
ronmental samples by gamma spectrometry, liquid scintillation
counting and alpha spectrometry: a review of analytical meth-
odology. J Environ Radioact 106:98–119
3. Burnett WC, Moon DS, Nour S, Horwitz P, Bond A (2003)
Preconcentration of radium isotopes from natural waters using
MnO2 resin. Appl Radiat Isot 59:255–262
4. Maxwell SL, Culligan BK (2012) Rapid determination of Ra-226
in environmental samples. J Radioanal Nucl Chem 293(1):
149–156
5. Burnett WC, Cable PH, Moser R (1995) Determination of Ra-228
in natural waters using extraction chromatography resins.
Radioactivity Radiochem 6(3):36–43
6. Nour S, Sharkawy AL, Burnett WC, Horwitz EP (2004) Radium-
228 determination of natural waters via concentration on man-
ganese dioxide and separation using Diphonix ion exchange
resin. Appl Radiat Isotopes 61:1173
7. Aleissa KA, Almasoud FL, Islam MS, L’Annunziata MF (2008)
Radium-228 analysis of natural waters by Cherenkov counting of
Actinium-228. Appl Radiat Isot 66:1954–1963
8. Maxwell SL (2006) Rapid method for 226Ra and 228Ra analysis in
water samples. J Radioanal Nucl Chem 270(3):651–655
9. Strelow FWE (1984) Distribution coefficients and ion exchange
behavior of 46 elements with a macroreticular cation exchange
resin in hydrochlroic acid. Anal Chem 56:1053–1056
10. Horwitz EP, McAlister D, Bond A, Barrans AB Jr (2005) Novel
extraction chromatographic resins based on tetraalkyldiglycola-
mides: characterization and potential applications. Solvent Extr
Ion Exch 23(3):319–344
11. Currie LA (1968) Limits for qualitative and quantitative deter-
mination. Anal Chem 40:586
Table 5 228Ra results with
interferences addedInterference added 90Sr/90Y 90Sr/90Y 238U 226Ra
5.92 Bq 29.6 Bq 3.7 Bq 4.81 Bq
1.046 Bq 228 Ra added 228Ra recovery
(%)
228Ra recovery
(%)
228Ra recovery
(%)
228Ra recovery
(%)
1 95.7 92.7 99.4 92.7
2 94.9 96.0 93.0 103.0
3 98.8 89.8 94.5 95.9
4 96.5 93.4 94.4 88.1
Avg. 228Ra Recovery (%) 96.5 93.0 95.3 94.9
SD 1.7 2.6 2.8 6.3
Table 6 Blank results with 1,000 pCi/L (37 Bq/L) 90Sr added
MDC
(pCi/
L)
Blank
sample
(pCi/L)
MDC
(mBq/
L)
Blank
sample
(mBq/L)
Decontamination
factor (Sr/Y)
1 0.20 0.19 7.35 7.02 5,267
2 0.21 0.25 7.85 9.13 4,054
3 0.19 0.25 7.18 9.35 3,956
4 0.21 0.32 7.64 11.68 3,166
5 0.20 0.21 7.26 7.86 4,707
Avg. 0.202 0.244 7.456 9.010 4,230
SD 0.008 0.048 0.280 1.771 797
% RSD 3.8 19.7 3.8 19.7 18.8
10 h count time
S. L. Maxwell et al.
123
Author's personal copy